How Often Is Chemotherapy Used for Eight Common Cancers? A Large International Study Finds Wide Differences Between Regions

Table of Contents

• Key Points
• Background: Why This Research Matters
• How the Study Was Conducted
• Key Findings: Chemotherapy Use by Cancer Type
• Age Differences in Chemotherapy Use
• Time to First Chemotherapy Treatment
• Chemotherapy Use in Stage 3 Colon Cancer
• What These Findings Mean for Patients
• Limitations: What the Study Could Not Determine
• Recommendations and Next Steps
• Frequently Asked Questions
• Source Information

Key Points

• In a study of nearly 800,000 patients in Norway, the UK, Canada, and Australia, chemotherapy use for eight common cancers varied widely between regions.
• For ovarian cancer, chemotherapy use ranged from about 48% to 81% of patients; for liver cancer, from about 4% to 51%.
• Patients aged 85 to 99 had 20 times lower odds of receiving chemotherapy than those aged 65 to 74, with variation between regions.
• Median time from diagnosis to first chemotherapy varied widely, especially for rectal cancer, where it was 89.2 days.
• The study could not adjust for stage, comorbidity, performance status, or treatment intent, and did not collect race or ethnicity data.

Background: Why This Research Matters

Chemotherapy is a key treatment for many patients with cancer. It is a type of systemic anti-cancer therapy, meaning it travels through the bloodstream to reach cancer cells throughout the body. Despite its importance, there was very little information comparing how often chemotherapy is used in different countries.

Previous studies had described differences in chemotherapy use between patient groups, but these studies were typically conducted within single countries. Before this study, researchers identified 18 primary studies that examined chemotherapy use across up to 12 regions. Most of these studies were within a single continent, and they covered only five of the cancer sites included in the current study: oesophageal, stomach, colon, rectal, and pancreatic cancer.

The earlier evidence did not formally measure the size of differences between regions. It also did not examine how long patients waited from diagnosis to their first chemotherapy treatment. This gap in knowledge mattered because time to first chemotherapy is especially important for treatments given before surgery (neo-adjuvant) or after surgery (adjuvant).

This study was part of the International Cancer Benchmarking Partnership (ICBP). The ICBP is a collaboration of clinicians, policy makers, researchers, and data experts in seven countries: Australia, Canada, Denmark, Ireland, New Zealand, Norway, and the United Kingdom. The partnership seeks to explain differences in cancer survival between high-income countries that have comprehensive cancer registries, similar health system spending, and universal health care.

How the Study Was Conducted

This was a population-based study, meaning it included all eligible patients in the participating regions rather than selecting a sample. The study used data from Norway and four UK nations: England, Northern Ireland, Scotland, and Wales. The study also used data from eight Canadian provinces: Alberta, British Columbia, Manitoba, Newfoundland and Labrador, Nova Scotia, Ontario, Prince Edward Island, and Saskatchewan. The study also used data from two Australian states: New South Wales and Victoria.

All participating regions had high-quality population-based cancer registries. Information about chemotherapy use came from cancer registry records linked to clinical or patient management system data, or to hospital administration data.

Who was included in the study:

• Patients aged 15 to 99 years at diagnosis
• Patients with a new primary diagnosis of one of eight cancers: oesophageal, stomach, colon, rectal, liver, pancreatic, lung, or ovarian cancer
• Patients diagnosed between Jan 1, 2012, and Dec 31, 2017 (with some variation in study periods between regions depending on data availability)
• Patients without another primary cancer diagnosis occurring within 5 years before to 1 year after the main cancer diagnosis, or during the study period
• Patients not registered solely from a death certificate
• Patients without cancer site-sex discordance (applicable to ovarian cancer only)

How chemotherapy use was defined: The researchers counted at least one administration of chemotherapy from 31 days before to 365 days after the date of diagnosis. Including a month of observation before the registered diagnosis date allowed the study to capture neo-adjuvant chemotherapy. Neo-adjuvant chemotherapy is treatment given before surgery. This chemotherapy was started before the final tissue-based diagnosis was confirmed.

The definition of chemotherapy aimed to include all types of systemic anticancer therapy, including immunotherapy. It covered all administration settings (such as hospital or community facility) and all routes (such as intravenous or oral). In most regions, all types, settings, and routes of chemotherapy were captured. However, in some regions, one or more of these components were not included in the data sources used.

Most patients receive chemotherapy of more than one type, in different settings, and through different routes. Therefore, the overall estimates of chemotherapy use are probably unaffected by missing one or more components of the definition. The exception was cancer sites where oral chemotherapy as a single treatment was an option during the study period and oral chemotherapy activity was not captured.

How the data were analyzed: The study used a federated or distributed analysis model. This means that standardized patient-level datasets were created and quality-checked in each participating region. Data were then analyzed within each region using centrally developed computer code. Only aggregate group-level data were shared with the central team, without transferring patient-level data.

Random-effects meta-analyses were used to measure differences between regions. The researchers calculated 95% prediction intervals (95% PIs), which show the range of outcomes that would be expected in similar regions not included in the analysis. They also reported tau values, which represent the standard deviation and serve as direct measures of how much variation exists between regions.

The primary analysis did not adjust for stage at diagnosis because information on stage was incomplete for 17.3% (135,206) of patients. However, the researchers also presented stage-adjusted results for regions with high stage completeness (more than 70%) for every individual cancer site.

Information was not collected on patient race or ethnicity. No comparable information was available on other medical conditions (comorbidity), performance status, or treatment intent.

Key Findings: Chemotherapy Use by Cancer Type

Between Jan 1, 2012, and Dec 31, 2017, there were 893,461 patients with a new diagnosis of one of the studied cancers. Of these, 111,569 (12.5%) did not meet the inclusion criteria, leaving 781,892 patients in the analysis.

The study found large differences between regions in chemotherapy use for all eight cancers studied. The 95% prediction intervals were wide for every cancer type, meaning that the range of chemotherapy use across similar regions is substantial.

Chemotherapy use by cancer type (pooled estimate and 95% prediction interval):

• Ovarian cancer: pooled estimate 66.4%, range 47.5% to 81.2%
• Oesophageal cancer: pooled estimate 47.2%, range 34.9% to 59.8%
• Rectal cancer: pooled estimate 40.8%, range 22.3% to 62.3%
• Stomach cancer: pooled estimate 39.6%, range 25.7% to 55.5%
• Pancreatic cancer: pooled estimate 34.1%, range 17.2% to 56.3%
• Lung cancer: pooled estimate 31.4%, range 17.9% to 49.0%
• Colon cancer: pooled estimate 29.7%, range 18.6% to 43.8%
• Liver cancer: pooled estimate 16.1%, range 3.5% to 50.7%

To put these numbers in perspective, the pooled estimate is the average across all participating regions. The prediction interval shows the range where 95% of similar regions would be expected to fall. For ovarian cancer, the pooled estimate of 66.4% means that about 2 in 3 patients received chemotherapy. The range of 47.5% to 81.2% means that in some regions, fewer than half of patients received chemotherapy, while in others, more than 4 in 5 did.

For liver cancer, the variation was especially striking. The pooled estimate was 16.1%, meaning about 1 in 6 patients received chemotherapy. But the range spanned from 3.5% (about 1 in 29 patients) to 50.7% (about 1 in 2 patients).

The researchers noted that they did not attempt to identify which regions had high or low chemotherapy use. They gave three reasons for this. First, there is no global agreement on the best amount of chemotherapy use in a population. Such amounts cannot be determined without information about other treatments and a range of patient characteristics. Second, small differences in how chemotherapy use was recorded might lead to large differences between specific regions, while having minimal effect on overall variation. Third, because the sample of regions is small, it cannot be assumed that the observed range represents the full range across all high-income countries with universal health care.

Age Differences in Chemotherapy Use

The study found large differences in chemotherapy use by age group. Patients aged 85 to 99 years had 20 times lower odds of receiving chemotherapy than those aged 65 to 74 years. The odds ratio was 0.05, with a 95% prediction interval of 0.01 to 0.19.

An odds ratio of 0.05 means that the odds of receiving chemotherapy were 95% lower for the oldest age group compared with the reference group. The wide prediction interval (0.01 to 0.19) indicates that this age difference varied substantially between regions.

The reference group for all age comparisons was patients aged 65 to 74 years, which was the most common age group in the analysis sample. The study also examined chemotherapy use in patients aged 15 to 64 years and 75 to 84 years compared with the 65 to 74 year group. These analyses were adjusted for sex, age, cancer site, and diagnosis year.

The researchers used categorical age groupings to aid interpretation and to simplify the federated analysis. There were no missing data in the analysis sample for age group, sex, cancer site, or diagnosis year.

Time to First Chemotherapy Treatment

The study found large variation between regions in the median time from diagnosis to first chemotherapy treatment. This variation differed by cancer site.

For rectal cancer, the median time to first chemotherapy was 89.2 days (about 3 months). The 95% prediction interval ranged from −15.5 days to 193.9 days. The negative value means that in some regions, the median time to first chemotherapy was before the registered diagnosis date. This can happen because the recorded diagnosis date sometimes follows the start of treatment when data flows are not perfectly synchronized.

For all eight cancers considered together, the study examined median time to first chemotherapy treatment. The analysis used observed values without adjustment for patient characteristics (casemix adjustment).

The study also examined differences in median time to first chemotherapy by sex and by age group. These analyses were adjusted for sex, age group, cancer site, and diagnosis year.

Patients aged 85 to 99 years had slightly shorter median time to first chemotherapy compared with those aged 65 to 74 years. This difference was consistent between regions: −3.7 days, with a 95% prediction interval of −7.6 to 0.1 days. The prediction interval crossing zero means that in some regions, the difference was not statistically significant.

Chemotherapy Use in Stage 3 Colon Cancer

The researchers specifically examined chemotherapy use in patients with stage 3 colon cancer. They chose this group because colon cancer had high completeness of information on cancer stage. The reasons for using chemotherapy in this patient group have been well established for a long time.

For patients with stage 3 colon cancer, the variation between regions was greater than for all patients with colon cancer. The pooled estimate was 60.1%, with a 95% prediction interval of 38.5% to 78.4%.

This means that for stage 3 colon cancer, about 3 in 5 patients received chemotherapy on average. But in some regions, fewer than 2 in 5 patients received it, while in others, more than 3 in 4 did.

The researchers also compared observed chemotherapy use against estimated population health need for chemotherapy by cancer site. They used estimates from evidence-based indications or comparative epidemiological data for this comparison.

What These Findings Mean for Patients

This study shows that where a patient lives may influence whether they receive chemotherapy and how quickly they start treatment. The differences between regions were large for all eight cancers studied.

The findings have several important implications:

• For patients: If you are diagnosed with one of these cancers, the likelihood that you will be offered chemotherapy may depend partly on the region where you receive care. This does not mean that one approach is right and another is wrong, but it does mean that practice patterns differ.
• For older patients: Patients aged 85 to 99 were much less likely to receive chemotherapy than younger patients. This difference was consistent across regions, but the size of the difference varied. The reasons for this age difference are not fully understood from this study.
• For timing of treatment: The time from diagnosis to first chemotherapy varied widely between regions, especially for rectal cancer. This variation could affect outcomes, particularly when chemotherapy is given before or after surgery.
• For health systems: The study highlights the need to understand why these differences exist. Possible explanations include differences in clinical guidelines and how they are put into practice, service organization, system capacity, health-care professional cultures, patient treatment preferences, and access to clinical trial protocols.

The researchers emphasized that the underlying reasons for these patterns need to be established to guide efforts to improve patient outcomes. They also noted that incorporating measures of chemotherapy use into international comparative survival studies and routine cancer surveillance is warranted.

Limitations: What the Study Could Not Determine

The study had several limitations that are important to understand:

• Missing stage information: Information on stage at diagnosis was incomplete for 17.3% (135,206) of patients. The primary analysis did not adjust for stage at diagnosis for this reason.
• No information on comorbidity or performance status: The study could not account for other medical conditions or how well patients were functioning, which can affect treatment decisions.
• No information on treatment intent: The researchers could not distinguish between chemotherapy given with curative intent (aiming to cure) versus palliative intent (aiming to relieve symptoms).
• Incomplete data in some regions: Chemotherapy use data were deemed incomplete in Newfoundland and Labrador (all sites) and Saskatchewan (ovarian cancer only). Estimates for Prince Edward Island for ovarian and liver cancer were not available because of small numbers of patients.
• Potential undercounting of oral chemotherapy: If oral chemotherapy was used as a single treatment, potential undercounting could substantially underestimate overall chemotherapy use for colon, rectal, and liver cancer, but not the other cancer sites.
• No information on race or ethnicity: The study did not collect information on patient race or ethnicity.
• Limited number of regions: Because the sample of regions is small, it cannot be assumed that the observed range represents the full range across all high-income countries with universal health care.
• Applicability: Results are likely to be applicable to regions with high human development index values and universal health-care coverage, such as those included in this study, but not to other settings.

The researchers also noted that they did not attempt to formally identify high or low chemotherapy use regions. They explained that there is no global consensus about the best amounts of chemotherapy use in a population. Such amounts cannot be inferred confidently without information about the use of other treatment methods and a range of patient characteristics.

Recommendations and Next Steps

The study authors made several recommendations based on their findings:

1. Understand the causes of variation: Future work should examine tumour, patient, and health-care factors that underlie chemotherapy use in different countries and health systems.
2. Examine health-care factors: These should include potential differences in clinical guidelines and their implementation, service organization, system capacity, health-care professional cultures, patient treatment preferences, and access to clinical trial protocols.
3. Include chemotherapy measures in survival studies: Incorporating measures of chemotherapy use in both international comparative survival studies and in routine cancer surveillance is warranted.
4. Address age disparities: The large and variable age group differences in chemotherapy use need to be better understood.
5. Investigate timing differences: The reasons for variation in time to first chemotherapy treatment, particularly for rectal cancer, should be explored.

For patients, the key message is that chemotherapy use and timing vary substantially between regions. If you have questions about your treatment options, including whether chemotherapy is appropriate for your situation and how soon it should start, discuss these with your cancer care team. The study findings do not indicate that any one approach is correct, but they do highlight the importance of understanding why practice patterns differ.

Frequently Asked Questions

How much does chemotherapy use vary between regions for these cancers?

In a study of nearly 800,000 patients across Norway, the UK, Canada, and Australia, chemotherapy use varied widely by region for eight common cancers. For ovarian cancer it ranged from about 48% to 81% of patients; for liver cancer from about 4% to 51%. The researchers did not identify which regions had high or low use.

How long might I wait from diagnosis to my first chemotherapy treatment?

The time from diagnosis to first chemotherapy varied widely between regions and by cancer type. For rectal cancer, the median time was 89.2 days, about three months, with a prediction interval from −15.5 to 193.9 days. A negative value means that in some regions the median time was before the registered diagnosis date.

What does it mean if a cancer has a pooled chemotherapy estimate of 66.4%?

The pooled estimate is the average across all participating regions. For ovarian cancer, 66.4% means about 2 in 3 patients received chemotherapy. The prediction interval of 47.5% to 81.2% shows the range expected in similar regions: in some, fewer than half received it; in others, more than 4 in 5 did.

Did the study look at chemotherapy for stage 3 colon cancer specifically?

Yes. For stage 3 colon cancer, the pooled estimate was 60.1%, with a prediction interval of 38.5% to 78.4%. This means about 3 in 5 patients received chemotherapy on average, but in some regions fewer than 2 in 5 did, while in others more than 3 in 4 did. Variation was greater than for all colon cancer patients.

What could the study not determine about chemotherapy use?

The study could not adjust for stage at diagnosis because stage information was incomplete for 17.3% of patients. It also had no information on other medical conditions, performance status, treatment intent, race, or ethnicity. Chemotherapy data were incomplete in some regions, and oral chemotherapy may have been undercounted for colon, rectal, and liver cancer.

What should I do if I have questions about chemotherapy for my cancer?

If you have questions about your treatment options, including whether chemotherapy is appropriate for your situation and how soon it should start, discuss these with your cancer care team. The findings do not indicate that any one approach is correct, but they do highlight that practice patterns differ between regions and the reasons need to be understood.

If I'm diagnosed with ovarian, liver, or rectal cancer, when should I seek a second opinion about chemotherapy?

Chemotherapy use varies widely between regions for these cancers. For ovarian cancer, about 48% to 81% of patients received chemotherapy depending on the region; for liver cancer, about 4% to 51%; for rectal cancer, about 22% to 62%. Time from diagnosis to first chemotherapy also varied, with rectal cancer's median around 89 days. Because practice patterns differ, a second opinion can help clarify whether chemotherapy is appropriate for your situation and how soon it should start. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: Use of chemotherapy in patients with oesophageal, stomach, colon, rectal, liver, pancreatic, lung, and ovarian cancer: an International Cancer Benchmarking Partnership (ICBP) population-based study

Authors: Sean McPhail, Matthew Barclay, Shane A. Johnson, Ruth Swann, Riaz Alvi, Andriana Barisic, et al.

Publication: Lancet Oncology 2024; 25: 338–51

Funding: International Cancer Benchmarking Partnership (funded by the Canadian Partnership Against Cancer, Cancer Council Victoria, Cancer Institute New South Wales, Cancer Research UK, Danish Cancer Society, National Cancer Registry Ireland, The Cancer Society of New Zealand, National Health Service England, Norwegian Cancer Society, Public Health Agency Northern Ireland on behalf of the Northern Ireland Cancer Registry, DG Health and Social Care Scottish Government, Western Australia Department of Health, and Public Health Wales NHS Trust).

Note: This patient-friendly article is based on peer-reviewed research published in Lancet Oncology. The original article is an Open Access article distributed under the terms of the Creative Commons CC BY 4.0 license.